PRAIRIE ESTABLISHMENT IN AGRICULTURAL SOILS

Katie Simonsen, E. Howell and Josh Posner

INTRODUCTION

This project, initiated in 1999, had two objectives: 1) to learn how to reestablish prairie on rich agricultural soils; and 2) to establish a natural “check” for environment variables (e.g soil health variables, CO2 sequestration) in the heart of the core cropping systems trial.  The experimental variables for prairie establishment were high and low density seed mixes.  To help “link” the prairie plots to the cropping systems plots established nine years earlier, a continuous corn treatment was also randomized among the prairie seeding treatments. 

METHODOLOGY

The WICST prairie experimental design

The experimental design was a randomized complete block with three replicates of three treatments established in a region of WICST previously set aside as a grass waterway (Fig. 1). All plots measured 510 feet long as the core trial, however, the width was varied to fit within the constraints of the existing agricultural trial; the two western block plots are 45 feet wide while the eastern block plots are 40 feet wide except plot 501, which had to be 45’ wide to accommodate the 6-row corn planter.

The three treatments used in this study were two prairie seed mixes of differing species composition and a cropping system control planted as continuous corn (CC).  The high diversity (HD) prairie seed mix has 25 species, the low diversity (LD) prairie seed mix has six planted species (see Table 1 for species planted and seeding rate).

The WICST prairie seed mix description 

The low diversity (LD) seed mix has a total of six species from three guilds (legume, C4 grass, and summer forb) and three families (Poaceae, Fabaceae, and Asteraceae) (Table 1). The guilds, or functional groups were based on Kindscher and Wells (1995) who described and studied 18 guilds on the Konza Prairie in Kansas.  The three guilds were chosen because they are also represented in the main WICST cropping systems, and because it is similar to the CP-2 option for grassland restoration offered by the Conservation Reserve Program (CRP).  The LD seed mix is currently selling for $450/acre.

These same six species in the LD mix, along with 19 additional species, are present in the high diversity (HD) mix for a total of 25 species from eight guilds (Table 1). These 25 species represent eleven families (Apiaceae, Asclepiadaceae, Asteraceae, Commelinaceae, Fabaceae, Lamiaceae, Liliaceae, Poaceae, Rosaceae, Rubiaceae, and Violaceae).  The eight guilds in the HD treatment were chosen to represent the eight guilds presented in Kindscher and Wells, and to include species with a range of bloom times and structural characteristics.  This treatment is modeled after the CP-25 seed mix of the CRP.  The current average cost for the HD seed mix is $1200/acre.  Both mixes are quite expensive partly because they have been well cleaned.

Planting rates and species description: 

Species were planted in rates that reflect not only their importance in the community (grasses are planted at a much higher rate than forbs) but also their relative seed weights (Table 1). For example, Potentilla arguta (prairie cinquefoil) has incredibly fine seeds and the approximate seed count per ounce is much higher, and hence the seeding rate is much lower, than that of Rosa Carolina (wild rose) which has large, heavy seeds.  In addition to seed weight, the planting rates accounted for the aggressiveness of the some of the species.  In general, those species that easily establish and can dominate a planting, such as Helianthus laetiflorus (showy sunflower) and Aster novae-angliae (New England aster) were seeded at a lower rate than some of the more conservative species like Asclepias tuberosa (butterfly weed) or Allium cernuum (wild onion). 

*Only 2 species in the HD mix not yet established.

Site preparation and planting

The site to be used in this study was plowed and planted to soybeans for the 1998 growing season.  After the bean harvest the land was chisel plowed several times and treated with glyphosate at the standard concentration of 3%.  To further attempt to reduce weed pressure, a tillage pass and herbicide treatment was applied two weeks before planting and a final tillage pass done a few days before planting.  The legumes were first scarified, and all species had been cold-treated over the preceding winter. On June 4, 1999 the seeds were hand-broadcasted onto the prepared soil after being separated into light-seeded forb, heavy-seeded forbs, and grasses for more even planting.  About half of the seeds were used in the first pass, and the rest were used with a pass perpendicular to the first.  The plots were then cultipacked after the broadcast was complete.  Weeds however quickly dominated the prairie plot (primarily lambsquarter and pigweed) and were clipped to allow some prairie seed germination.  In the third year after establishment (June, 2002) the plots were again clipped and stinger was spot sprayed to control thistles in the prairie plots.  In the spring of 2003 (4/23/03) the prairie plots were burned. In June of 2003, prairie plots were again spot sprayed for thistles but at 1/5 the amount of Stinger that was applied last year suggesting that we have the thistle population declining.

Table 2 is the agronomic diary of activities done on the corn “check plots” that were planted at the same time as  the prairie plots.  Table 5 is a summary of fall nitrate levels on the prairie plots as well as the WICST continuous corn plots (CS1).

Table 2.  Corn check plots in the prairie planting: Input/Output Data - Arlington

¹ outside 6 rows sprayed on 11/6/01 to control quackgrass invasion and dandelions

* plot 501 received no N per PPNT

Sampling

Permanent meter-square quadrats: 

Each plot was divided widthwise into four equal sections (see Fig. 2).  The location of a single, one-meter-square quadrat (1m x 1m) was randomly assigned to each section.  These permanent quadrats were marked with a wooden stake in the northwest corner and GPS coordinates of each quadrat were recorded in the spring of 2000 and are listed at the end of this report.

Before the prairie plots were seeded, meter-square plastic sheets were placed over the permanent quadrats to prevent seeds from landing in them.  Each permanent quadrat was later planted with pre-counted seed packets so the exact number of seeds of each species was known and equal to the planting rates in the rest of the field.

Sub-plot quarter meter-square quadrats:  

Prior to each sampling date, additional quadrats (1/2m x 1/2m) were identified .  Each plot had a total of 20 subplot quadrats; five in each stratified section. 

Sampling dates

There were two to three sampling sessions each year and findings over the year are found in Table 3.  A sampling session refers to the effort to sample all permanent and sub-plot quadrats to record either presence or absence of each planted species.

All species rooted in each quadrat were identified and recorded as present. Frequency was determined by dividing the number of quadrats in which a species was present by the total number of quadrats (Table 4).

Eye level estimation, meaning the cover of each category, was recorded by visual estimation as that which is visible from the eye level of the observer (approx. 150 cm).  So, for instance, if a tree held most of its biomass above 150 cm, the cover of the trunk would be estimated, but the canopy cover would not be.  The categories used in this study were total cover, graminoid cover, forb cover, native cover, non-native cover.  The graminoid and forb cover, when combined, equal total cover, as does the native and weed cover when combined.  Not all categories were used in each session

Qualitative walk-throughs

 In addition to quantitative measurements, a qualitative walk-through was used to account for species that were present but not recorded in the quadrat sampling. This was done by making a zigzag pattern through each section of each of the three treatments for fifteen minutes and recording species presence and abundance.  Abundance was noted by a scale of one to six with one being very rare and six being dominant and was assigned after each section was complete.

RESULTS

In Table 4 we have a summary of the presence or absence of the planted species.  The first three columns focus on the low diversity mix, and already during the sampling in 1999, all six species were observed either in the quadrats, or during the walkthroughs.  However, Baptisia Leucantha subsequently disappeared from the low-density plots.  With the high diversity mix, not surprisingly the grasses came in immediately, and by year 2 there were eighteen species identified.  By year 3, twenty-one of the twenty-five species planted had emerged, either in the quadrats (17) or only observed in the walkthrough (4).  It was a pleasant surprise that some of the more slow-growing or conservative species (e.g. Allium cernuum, Asclepias tuberosa, Astragalus Canadensis, Silphium integrifolium and Viola pedatifida) were seen in the high diversity plots by the second year of monitoring.

Table 4. Presence of planted species in the two prairie treatments for the three years of the study.

As can be seen in Table 2, corn yields in the prairie plots were good between 1999 and 2002 and quite similar to the continuous corn plots (CS1) in the Cropping Systems Trial (established in 1990). 

Fall nitrates were also monitored and as can be seen from Table 5, the prairie corn and the WICST CS1 corn were similar in value and by 2001, already a good deal higher than in the prairie plots.  Having permanent ground cover in the established prairie with species that have a deep-rooted or fibrous root system help prevent the buildup and potential leaching of soil nitrate-N.

CONCLUDING REMARKS

After four years, the prairie restoration on rich agricultural soils has proven to be a success.  Very few annual weeds exist (lambsquarter, pigweed, horseweed, etc) anymore, and nearly all of the 25 species in the high diversity mix have been established.  Only Baptisia lecucantha (White wild indigo) did not persist in either of the mixes.  Continued monitoring of exotic weeds and periodic burning will be useful to the native prairie species.  The differences between the high and low diversity treatments will likely disappear as the plants and seeds spread over the years.  The 24 permanent quadrats (4/plot) will continue to be monitored to trace the changing species mix on the plots.  Not surprisingly, after prairie has been established, fall nitrates beneath the prairie have been drawn down to around 20-40 lbs N/3 ft a, which is considered a natural background level (mostly due to plant and animal decay) on undisturbed soils.

The prairie corn plots, although only 4 years into corn, are producing yields nearly equivalent to those of the continuous corn system in the surrounding WICST plots.  As a result, we now have an ecological ‘check’ to compare agricultural systems against.

Home | Who we are | Research | Publications | Links | Site map | Contact us

University of Wisconsin-Madison
College of Agricultural and Life Sciences
If you have any difficulty accessing this page, please contact our web master.

Site hosted by UW-Madison Center for Integrated Agricultural Systems